topic Re: SQL code against tables with '>' in headers suddenly failing? in Data Engineering

SQL code against tables with '>' in headers suddenly failing?

Kayl669 — Tue, 30 Apr 2024 09:53:10 GMT

Just want to post this issue we're experiencing here in case other people are facing something similar. Below is the wording of the support ticket request I've raised:

SQL code that has been working is suddenly failing due to syntax errors today. There have been no changes to the Databricks runtime we use (12.2).

The SQL code that is failing interrogates tables with '>' in the field names. The error is produced when you query these tables in a subquery or CTE and then SELECT from the subquery or CTE. Our code uses backticks to enclose these field names already.

Sample error message:
[PARSE_SYNTAX_ERROR] Syntax error at or near '>'.(line 1, pos 5) == SQL == event>eventData>identityState>customerId -----^^^

Re: SQL code against tables with '>' in headers suddenly failing?

Aviral-Bhardwaj — Tue, 30 Apr 2024 12:56:11 GMT

Please provide query here, query is not visible in the question

Re: SQL code against tables with '>' in headers suddenly failing?

Kayl669 — Tue, 30 Apr 2024 13:23:11 GMT

Code in the attached .png. Can you see this? Thanks

Re: SQL code against tables with '>' in headers suddenly failing?

Aviral-Bhardwaj — Tue, 30 Apr 2024 13:31:12 GMT

please dont use back ticks use and whole statement should not be in `` use ' ' for value only

Re: SQL code against tables with '>' in headers suddenly failing?

Kayl669 — Tue, 30 Apr 2024 13:37:40 GMT

This was code that was working yesterday? You have to use backticks because the field names of the table contain characters akin to spaces?

Re: SQL code against tables with '>' in headers suddenly failing?

Kayl669 — Fri, 03 May 2024 07:55:44 GMT

The point that we've got to with this is that MS Support / Databricks have acknowledged that they did something and are working on a fix. "The issue occurred due to the regression in the recent DBR maintenance release...Our engineering team is working on a hotfix for the issue, which should be implemented by tomorrow EOD"